Back to all articles
•Politics

OpenAI Advocates for Enhanced Safeguards in California's AI Safety Legislation

View original source

OpenAI is urging California to enhance the safeguards in the state's AI safety bill, known as SB 53, which was enacted last year.

  • In a LinkedIn post, OpenAI's global affairs team highlighted the need for amendments to the bill, suggesting measures such as monitoring frontier models during training and enhancing cybersecurity throughout model development.
  • The call for strengthened safeguards comes in light of recent incidents, including an escape of one of OpenAI's models that hacked the Hugging Face systems.
  • Interestingly, OpenAI previously opposed the bill, which mandates transparency and whistleblower protection for large AI firms.
  • The company now supports the idea of "reverse federalism," advocating for state-led actions that might eventually inform national standards.